Munich Datageeks - February (March) 2016 Edition

It's time for our next Meetup. This time the Munich Network will host our Meetup at the Innovationswerk near the Ostbahnhof in Munich.

Please find the location plan for the Innovationswerk here:

Innovationswerk (https://www.google.de/maps/place/Munich+Network-Netzwerk+M%C3%BCnchen+e.V./@48.1248868,11.6029313,16.5z/data=!4m5!1m2!2m1!1sinnovation+werk+munich!3m1!1s0x0000000000000000:0x530459e9b75965e6)

Format:

• 2 presentations (each ca. 30-40 min incl. discussion)
• Of course time for networking + food + drinks before, in between and especially after the presentations
• Talks are held in English

The line up:

David Anderson -
Sketches of streaming data for fast, approximate big data analysis

Bio:

David Anderson has been a Senior Research Scientist, Chief Architect, and CTO. He leads the Big Data Engineering program at Data Science Retreat, which offers training in Scala, Spark, big data, and machine learning.
In recent years, David has worked on search engines, predictive analytics, recommender systems, and visualization products for companies throughout Europe. Previously he worked as a researcher, applying machine learning techniques to create better user experiences -- for toys, learning environments, and logistics optimization.

Abstract:

Busy web applications and sensor networks are creating new challenges for big data analytics. The scale involved can render straightforward (brute force) solutions infeasible. Sketches are a family of algorithms that can efficiently compute reasonable approximations of summary statistics on large datasets.

I plan to present a general overview of the topic, and then present one technique in depth -- hyperloglog -- which is an algorithm for cardinality estimation. Yes, that's right, in big data we spend our time coming up with complicated ways to count things.

Karim Jedda - Datamining the next series to watch

click here (http://funnybretzel.com/blog/datamining-the-next-series-to-watch-part-1/)